GPS Interference Testing

mP1 September 5, 2013 at 1:27am

@Chad

So why dont any copter manuf sell shielding made of aluminium for their controllers ? Why does nobody make a dish or similar shape to block vid tx from the controller (aka APM) ?

Phillip Jones July 8, 2012 at 2:24pm

The SNR measurement will give you much better information about the quality of the signal. If it's in the low 20's, you'll get lock but it's somewhat unreliable. Mid 30's and up is quite reliable.

Chad Frazer July 8, 2012 at 9:43am

Aluminum Foil is a notorious electromagnetic shielding agent, you can stand next to radar towers with your camera cloaked in one and it will still work (digital cameras) without it they sometimes wont even turn on!

Brushless motor, big time radio frequency generator. Some of the current flowing into the motor goes to generating a magnetic field that interacts with the permanent magnets glued to your motor case (depending on what type of motor you have or course). Some of that current gets emitted as waste heat. Some of that current gets thrown out with a certain value of energy which puts the emitted particle flat dab square in the GPS radio frequency band. There is your interference.

2.4 GHz signal. "All satellites broadcast at the same two frequencies, 1.57542 GHz (L1 signal) and 1.2276 GHz (L2 signal)." from Wikipedia. That second frequency, 1.22Ghz is very close to 2.4 Ghz, its double the 2.4 Ghz. That means that if the antenna is tuned to receive 1.2Ghz signal it will be prone to also have effects (see prior post by me on what atoms do about absorbing energy) by close and powerful sources of 2.4Ghz signals. The 1.2 antenna LIKES, but not LOVES that energy packet range.

Could you try carbon fiber of commonly used thicknesses, balsa wood, monokote, copper, water and perhaps spray paint of any kind?

Chad Frazer July 8, 2012 at 9:23am

Material Science time. For those of you who don't want to know this, here you go:

On an atomic scale, you can predict which materials will have an effect by realizing two basic truths about way our universe works.

First truth, we live in a digital type of universe, more commonly called Quanta or Quantum. Big meaningful word for the fact that particles have a requirement before they will interact with another particle. Big picture here, your oven leaks energy (heat) because the atoms that the oven walls are made of do not ABSORB the energy your heating elements put out. Atoms will only interact with a certain range of other atoms when it comes to absorbing or emitting energy. SO, if you know the energy level of the particles of your radio emitter (wavelength of the RF transmission, like 2.4 Ghz) then you can look up what atoms readily interact with that energy spectrum. Easy peasy right? A good example of this are certain calculations done for nuclear reactors when it comes to predicting how much radiation will leak from a reactor at power. I kid you not, matter really does behave like this.

The second thing you need to realize is how atoms arrange themselves in a given material. You can have the right atoms in a bucket but not get the desired effect until you properly assemble them. The geometric shapes that atoms arrange themselves in is of great significance to the end effect. That is where the bear lives, so to speak. What your talking about on this second point are the formulas and the manufacturing procedures that nations label top secret and has hopes that other nations, like say Iran will not be able to decipher...the "stealth" in the paint on that downed Sentinal UAV.

Mike Schwier July 7, 2012 at 9:38pm

The little blue bar that appears next to the altitude ticker tape is the Rate of Climb according to the Planner Manual. http://code.google.com/p/ardupilot-mega/wiki/MPGCS This may briefly appear when a 3d fix is acquired because the altitude is being corrected.

I dont agree that satellite counts reported by the GPS receiver is a good measure of the ability to use those satellites for accurate navigational purposes. The GPS receiver is only reporting the number of satellites it is able to receive a signal from, not whether it has successfully received data from the satellite and is able to use the satellite as part of the navigation solution. There is some good information in the post here: http://diydrones.com/forum/topics/gps-2d-vs-3d-fix?xg_source=activi... Make sure to read Ken's comments on the second page. There is also a satellite predictor tool referenced. http://www.navcomtech.com/Support/Tools/satellitepredictor/main.cfm

HDOP (Acronym for horizontal dilution of precision) would be a much better way to test how various materials affect the GPS. HDOP is a measure of the geometric quality of a GPS satellite configuration in the sky. HDOP is a factor in determining the relative accuracy of a horizontal position. The smaller the DOP number, the better the geometry.

I also noticed that you were testing on the ground and both of you were sitting near the aircraft. Your body could have been blocking reception of a satellite at any time. A better method would be to put the A/C on top of a camera tripod and sit under it. This should help to minimize the effects of your bodies and other nearby objects on GPS reception and your test results.

Keep up the good work. I look forward to the next video!

T3

Ted Van Slyck July 7, 2012 at 8:57pm

I use both Arducopter and Mikrokopter products. With Arducopter I have never read more than 9 sats. With Mikrokopter I have a shielded GPS and routinely hit 12 sats. The position hold feature is incredibly more accurate. My gut tells me that the APM would benefit greatly from creating a shield the the GPS.

Dr Mike Black July 7, 2012 at 3:47pm

Gentlmen,

It sounds to me as though we are doing similar tests.

I have to be honest, i am very much a noob on the APM2 but have extensive knowledge and development of the original AP (so the basis is there).

We are having a similar problem with the MTK, as in it does what it wants to do, and have tried all power configurations (Jumper and power settings/types), with varying degrees of failure. (It will lock up sometimes.....sometimes!)

With the APM2 built in MTK , which has been giving me nothing but problems, so I want to test with another GPS, see if its my set up or the GPS.

Based on my tests i have deduced the following.

Hooked Ublox onto the same port (right or wrong, i could find no elegant way of shutting down the on board GPS. So hooked them in parrallel.

30 sec Lock...no loss of lock, there are some parts of the data missing, if you look at the HUD you can see that no sats being shown, and HDOP is not registering..also another wierdness, that the IMU is moving around slightly in all directions when ublox connected. (we will come back to this later!)

Went back to MTK..nope....locks, loses lock...

So in the interestes of science, i flew 5 modes...FBW/Manual/Auto/circle/loiter to make sure all is well with the basics, then loaded some WP's hit auto....with MTK...no attempt at all to navigate..in fact the aircraft just circled downwind, until i got the willies and brought it home..(however it did not try to commit suicide, which is always a bonus!!)

connected ublox, WE HAVE NAVIGATION...its vague...lackluster, but its trying to track (PID's clearly need playing with) BUT...its trying...

So went back to MTK..locked..."Ah here we go!". got airbourn....tried to go to WP1...then half way there....it returns to the 'im lost but dont want to crash' mode again....

Again, landed, re-hooked ublox.....off she went again, tracked, no loss of lock...

So conclusion....in my case, (and many other from what i have read) MTK just is not cutting the mustard.

It would be helpful if any of the community has/is/or trying to do this, some input would be really great, if nothing else to clear up the vaguries of how the GPS sytem is implimented.

I have had problem after problem with the MTK units, either on their own, or in other applications over the last few years, IMHO they are too easy to blank/upset/annoy...never really experienced belligerence and or petulance in an inanimate object before : ) Perhaps dinner and flowers might be required to get them to put out for us!? )

Personally, i have no issue if a unit is screwey, but it must be realiably screwey. so you can create a work-around...not what seems to be a completely random set of behaviours, based upon, whether you have black or grey underware on, which type of car is parked outside no.30 on a wednesday morning, and if planet X is having a solar eclipse..

MTK units have given me issues on AP (all 6 we have flying) APM1 (2 off) Remzibi OSD (3 off) drops out all the time, loses home....

I am wondering if the unit itself, has enviromental requirements, that MUST be met, but only work fro a portion of this community, location depending.

Either way i have 8 of these suckers....and not one, behaves in a predictable fashion, in relation to itself and the others.

Ublox.....every time....on the nail! No questions...no mucking about....

Another issue here in the UK is that there are no WAAS or similar so we cant grab extra data from other sources..

So, what is the 'Correct procedure' for disabling the MTK and hooking the Ublox or any other GPS to the APM2 with the corrected data displayed on the GCS.

I strongly suspect that, what you guys are experiencing, is what i am also, for all the same reasons. Largly i have conducted similar tests as you with similar results...I can only default to the GPS itself..

Anyway looking forward to your thoughts on this issue.

Kind regards,

Mike.

Developer

Ryan Beall July 7, 2012 at 12:01pm

Just to add a little more emphasis on hdop: I pay attention to sats on the map page, but I mostly pay attention to hdop. For the Mtek gps, I didn't feel comfortable launching until I saw below 200 for hdop. When you don't have a lock hdop =9999. Then the more sats you get, it will start to drop. You can have a 3d lock with 6 sats and an hdop of 4-500 (wich is really quite bad) for a minute or so after you've gotten 3d lock. Ideally you want to see around 175 for hdop on a Mtek gps. I use that as my sanity check on a average solid lock gps day.

Hopefully this will make things a little more tangible for you!

P.S. adding HDOP on the map page was my contribution ;)

Developer

Ryan Beall July 7, 2012 at 11:54am

I've noticed the same issue with the Mtek gps not giving a 3d fix with 4 sats. It's really not an issue I guess because I wouldn't fly with less than 6 in most cases anyway. With that being said, APM is reporting the number of sats that the gps is reporting it currently has a lock on. The sat count can also be seen in the bottom left corner of the map (Sats:) you can also see the horizontal precision (hdop) there as well. This number will actually give you a better metric for measuring the accuracy of the gps. Mtek reports hdop (horizontal dilution of precision) aka +- how far you can expect to be off horizontally in meters. So you may have the same sat count for different materials but you can for sure see how much interference you have by watching your hdop.

A second note: You mentioned the blue bar popping up when you got a 3d fix and how that may be tied to gps. The blue bar is the VSI or vertical speed indicator. You should see this go up and down by lifting the plane up and down a few meters. It uses the barometric altitude to give you a climb or descent rate. With that being said the barometric altitude is slaved to the gps altitude to cancel any drift in the pressure altitude due to temperature etc. So when you got a lock on the gps, the "gps reported altitude" quickly resynced the baro altitude and caused a quick artificial indicated climb. Hope this helps: -Beall

Blogs

GPS Interference Testing

Comments